An ever-increasing reliance on information and the computing systems that produce, process, distribute, and maintain such information in its various forms, continues to put great demands on techniques for storing, backing up, and restoring such information. As businesses adopt techniques for centralizing information resources across an enterprise to enable collaboration and document management, storage, backup and restoration of such information becomes even more critical to the functioning of the enterprise.
In a typical enterprise-level collaboration and document-management platform, data from one or more user-oriented application objects can be stored in a centralized storage object such as a database. The database maintaining the data from the variety of application objects can be configured to maintain data relationships, or hierarchies, imposed by the application objects. The database can also be configured to store multiple versions of data objects, for example, documents and list items. Any backup and restoration scheme for such a database must maintain the relationships and hierarchies of the data created by the application objects, as well as be able to manage restoration of data object versions.
Traditional methods of backing up and restoring such a database are both personnel and resource intensive. For example, given that there are a multitude of application object-based areas in the database, a backup administrator would be responsible for knowing and selecting those portions of the database that are necessary to protect. Such an up-front, granular approach to data protection requires that the backup administrator, who may not be the database administrator, know the topology of the database in order to build a protection schema for a particular application object. Further, any changes made to the hierarchical object structure for an application would need to be taken into account by the backup administrator upon occurrence of those changes. This traditional backup approach is called a granular backup scheme, because the areas in the database associated with each application object are backed up separately. Under a granular backup scheme, restoration of data related to a particular application object can be reasonably targeted because each application object has its own set of backups.
An alternate method of backup is to backup the entire database monolithically. A traditional method of restoring data from such a monolithic backup of the database is to restore the entire database snapshot to a temporary area and then select the desired information from the database to include in an active target database. Drawbacks of such a back-end method of data selection are that it takes time to restore all the data from a database backup from which the desired information would then be selected and disk resources are consumed by the temporary copy of the database.
It is therefore desirable to have a mechanism that realizes the efficiency of monolithically backing up the entire database, thereby not requiring that a backup administrator be familiar with the structure of the database being backed up. It is further desirable that a user be able to selectively restore portions of the backed up database when desired so as to speed recovery time. It is also desirable that the restoration mechanism be capable of distinguishing those backed up objects that are versioned and to restore those versions in a manner compatible with normal operation of the collaboration and document-management platform.